Paskirstytos maišos lentelės modeliai ir jų saugumo problemos yra aktyviai nagrinėjama sritis akademinėje bendruomenėje. Šiame darbe keliama hipotezė, kad praktinės DHT tinklų realizacijos neužtikrina teoriniuose modeliuose priimtos sąlygos, kad mazgų prisijungiančių prie tinklo identifikatoriai bus generuojami atsitiktinai. Randamas atakų sudėtingumo įvertis, kai atsitiktinių identifikatorių generavimas yra privalomas ir kai identifikatorius galima pasirinkti laisvai. Hipotezės patvirtinimui atliekamas eksperimentas. Surenkami ir analizuojami vieno didžiausių DHT tinklų (BitTorrent DHT) duomenys. Aprašomas literatūroje siūlomas problemos sprendimo būdas ir praktinės problemos kylančios jį realizuojant. Pasiūlomas naujas praktiškai pritaikomas ir našus identifikatorių generavimo patikrinimo metodas bei metodas naujų identifikatorių generavimo greičiui DHT tinkle valdyti. / Distributed Hash Table models and its security implications, has long been a subject of interest. This thesis is based on the assumption that practical implementations do not enforce random node id generation regardless of the fact that theoretical models require node ids to be chosen by random and distributed in the address space uniformly. To measure the impact on the DHT network security if the assumption holds an analysis of attack complexity in both cases is performed. Results indicate that the complexity grows from O(1) to O(M) if the node id cannot be selected arbitrarily (M is the number of nodes in DHT network). Stated assumption is confirmed by analysing classic node id protection methods and performing analysis of BitTorrent DHT network. The reason for the lack of node id protection in practice is considered to be the complexity and performance penalty of the classic methods. To facilitate the implementations of DHT networks a new method to ensure random node id generation and copy protection is provided. Proposed method utilizes MACs based on shared keys to provide a proof of the ownership of the node id while still providing means to protect it from being copied. Efficiency of the proposed method is evaluated by conducting an experiment. In order to protect small DHT networks against a Sybil attack a method to control the speed of node id generation is also proposed.
Identifer | oai:union.ndltd.org:LABT_ETD/oai:elaba.lt:LT-eLABa-0001:E.02~2012~D_20131105_093701-72131 |
Date | 05 November 2013 |
Creators | Kriukas, Julius |
Contributors | Kavaliūnas, Rimantas, Blažauskas, Tomas, Kaunas University of Technology |
Publisher | Lithuanian Academic Libraries Network (LABT), Kaunas University of Technology |
Source Sets | Lithuanian ETD submission system |
Language | Lithuanian |
Detected Language | English |
Type | Master thesis |
Format | application/pdf |
Source | http://vddb.library.lt/obj/LT-eLABa-0001:E.02~2012~D_20131105_093701-72131 |
Rights | Unrestricted |
Page generated in 0.0026 seconds